This document discusses an artificial immune system approach for detecting computer viruses. It begins by providing background on artificial immune systems and how they can be applied to computer security similar to how the human immune system distinguishes self from non-self. It then describes the proposed artificial immune system-based virus detection system, which includes a signature extractor that generates signatures for non-self programs that do not match self programs, and a signature selector that analyzes the signatures to determine if they belong to viruses or self programs. The system aims to detect unknown viruses through an adaptive process of learning virus signatures.